Overall diagnostic accuracy of different MR imaging sequences for detection of dysplastic nodules: a systematic review and meta-analysis

To assess the overall diagnostic accuracy of different MR imaging sequences in the detection of the dysplastic nodule (DN).

PubMed, Cochrane Library, and Web of Science were systematically searched. Study selection and data extraction were conducted by two authors independently. Quality assessment of diagnostic accuracy studies (QUADAS) 2 in RevMan software was used to score the included studies and assess their methodological quality. A random-effects model was used for statistical pooling by Meta-Disc. Subgroup analysis and sensitivity analysis were used to explore potential sources of heterogeneity.

Fourteen studies (335 DN lesions in total) were included in our meta-analysis. The area under the curve (AUC) of summary receiver operating characteristic (SROC) of T₂WI was 0.87. Pooled sensitivity, specificity, positive likelihood ratio (PLR), and negative likelihood ratio (NLR) of DWI were 0.81 (95%CI, 0.73–0.87), 0.90 (95%CI, 0.86–0.93), 7.04 (95%CI, 4.49–11.04), and 0.24 (95%CI, 0.17–0.33) respectively. In the arterial phase, pooled sensitivity, specificity, PLR, and NLR were 0.89 (0.84–0.93), 0.75 (0.72–0.79), 3.72 (2.51–5.51), and 0.17 (0.12–0.25), respectively. Pooled sensitivity, specificity, PLR, and NLR of the delayed phase were 0.78 (0.72–0.83), 0.60 (0.55–0.65), 2.19 (1.55–3.10), and 0.36 (0.23–0.55) separately. Pooled sensitivity, specificity, PLR, and NLR of the hepatobiliary phase were 0.77 (0.71–0.82), 0.92 (0.89–0.94), 8.74 (5.91–12.92), and 0.24 (0.14–0.41) respectively. Pooled sensitivity, specificity, and PLR were higher on DWI and hepatobiliary phase in diagnosing LGDN than HGDN.

MR sequences, particularly DWI, arterial phase, and hepatobiliary phase imaging demonstrate high diagnostic accuracy for DN.